Question who is most responsible for saving AMD

[positivedoppler]

It was only 5 years ago AMD ended the year 2.3 billion in debt, 780 million in cash, and under 4 billion in sale.
https://www.wsj.com/market-data/quotes/AMD/financials/annual/balance-sheet
I still recall the excessive doom and gloom posts from a few members here who can no longer be found.

The AMD BR wagon was only 5 years ago, and now they are back and once again a serious threat to Intel and Nvidia, who do you think was the most responsible for saving the ship? Lisa Su, Rory Read, Jim Keller, Papermaster, etc..?

 

[EXCellR8]

Why do you have to try and troll every AMD thread ?

That's just what delusional Intel loyalists do... they all mad rn

 

[guachi]

1. The Ryzen design team.
2. The higher ups who thought that focusing on Ryzen was the right thing to do.

 

[moinmoin]

Mark Papermaster for setting up the projects and staffing them (Zen team). He essentially built the base necessary for the turn around.
Lisa Su for first handling Semi Custom contacts (ensuring AMD could survive the "drought"), then making AMD much more focused once getting in CEO position, making best possible use of what AMD could offer.

 

[yeshua]

It was only 5 years ago AMD ended the year 2.3 billion in debt, 780 million in cash, and under 4 billion in sale.
https://www.wsj.com/market-data/quotes/AMD/financials/annual/balance-sheet
I still recall the excessive doom and gloom posts from a few members here who can no longer be found.

The AMD BR wagon was only 5 years ago, and now they are back and once again a serious threat to Intel and Nvidia, who do you think was the most responsible for saving the ship? Lisa Su, Rory Read, Jim Keller, Papermaster, etc..?

Talented engineers, solid management decisions (like getting rid of GloFo), investors, and Intel who has dragged their butts for so long. If Intel had 8-core Ice Lake/Tiger Lake CPUs running at 5GHz at 95W TDP right now, no one would be talking about AMD. And, no, AMD is not a threat to NVIDIA. It may look so for some people who love AMD to no end, but it's just an illusion.

Meanwhile people choose not to talk about their worst decision ever - which was to buy ATI. When did it happen? 15 years ago? They still don't have a single product which justified that purchase. The APU (the way it was initially envisaged) has never happened. This purchase almost tanked AMD.

 

[BigDaveX]

Talented engineers, solid management decisions (like getting rid of GloFo), investors, and Intel who has dragged their butts for so long. If Intel had 8-core Ice Lake/Tiger Lake CPUs running at 5GHz at 95W TDP right now, no one would be talking about AMD.

Yes they would, since AMD would still have more cores thanks to the chiplet design. But it would at least be a situation more comparable to the original Zen versus Kaby Lake (or possibly Coffee Lake v1) rather than what we have now.

Meanwhile people choose not to talk about their worst decision ever - which was to buy ATI. When did it happen? 15 years ago? They still don't have a single product which justified that purchase. The APU (the way it was initially envisaged) has never happened. This purchase almost tanked AMD.

It was a good decision in the longer run, since it saved them the trouble of developing their own IGPs in-house, and they raked in a huge amount from the PS4 and Xbox One deals (and I'm sure at least got in enough money from the Wii U to redecorate the executive lounge) at a time when their CPU line-up was floundering. Their real mistake was paying ridiculously over the odds, which ended up being doubly damaging considering Intel retook the performance lead right as AMD closed the deal.

 

[DrMrLordX]

Ultimately, the customers who kept buying their junk products until they finally got their butts in gear.

So . . . me?

Allegations that were not only never proven but no one from the Justice Department even looked into the matter because there was nothing there.

Wow, that's a first. A fan of Hector Ruinz. I applaud you for your unique tenacity in defending that. Uh. Guy.

 

[yeshua]

Yes they would, since AMD would still have more cores thanks to the chiplet design. But it would at least be a situation more comparable to the original Zen versus Kaby Lake (or possibly Coffee Lake v1) rather than what we have now.

And Intel surely cannot stick their dies together just like AMD does. Oh, wait, they had the Q6600 quad-core CPU around 15 years ago or so. In fact I believe Intel had the expertise to glue dies together before AMD even thought about that.

It was a good decision in the longer run, since it saved them the trouble of developing their own IGPs in-house, and they raked in a huge amount from the PS4 and Xbox One deals (and I'm sure at least got in enough money from the Wii U to redecorate the executive lounge) at a time when their CPU line-up was floundering. Their real mistake was paying ridiculously over the odds, which ended up being doubly damaging considering Intel retook the performance lead right as AMD closed the deal.

AMD could have easily licensed GPUs from ATI just like Intel recently licensed AMD's one (Kaby Lake-G). And that would have cost them a LOT less then buying out ATI. And that would have left them not with an insane amount of debt but enough cash to take on Intel.

AMD's GPU division is still nowhere near the level ATI was back when it was alive and kicking and RDNA success is not really a success considering they have reached performance/power parity with the competition which uses a twice worse node with extra transistors for RTRT and denoising/AI image upscaling.

Spoiler

Artem S. Tashkinov

 

[BigDaveX]

And Intel surely cannot stick their dies together just like AMD does. Oh, wait, they had the Q6600 quad-core CPU around 15 years ago or so. In fact I believe Intel had the expertise to glue dies together before AMD even thought about that.

Yes, whipping up multi-die CPU packages to counter AMD is one of the oldest things in Intel's playbook... and the fact that they didn't immediately counter Zen by just slapping two Kaby Lake dies onto a package to instantly make an 8C CPU indicates that Skylake and its offshoots don't have the necessary on-die logic to do so, and/or there are other issues that make it infeasible.

AMD could have easily licensed GPUs from ATI just like Intel recently licensed AMD's one (Kaby Lake-G). And that would have cost them a LOT less then buying out ATI. And that would have left them not with an insane amount of debt but enough cash to take on Intel.

It'd have been cheaper, yes... but it would have also left them vulnerable to Intel swooping in further down the line and making ATI an even better offer, leaving AMD up the creek unless they managed to get a deal with nVidia. Like I said above, buying out ATI wasn't a bad idea in principle, it was just that the timing and execution of the deal both went horribly wrong; a merger would probably have been a better idea than a full-on takeover, but Ruiz's ego got in the way.

AMD's GPU division is still nowhere near the level ATI was back when it was alive and kicking

Seems like it's been about the same to me - a few years where they were well ahead of nVidia, combined with years when they've been about equal, and others when they were clearly behind. ATI's solo years weren't all spent delivering the kind of beat-downs to nVidia they dealt out during the eras of the Radeon 9700/9800 - and had AMD not bought them out, then the Radeon 2900 would still have sucked in comparison to the GeForce 8 series.

 

[ksec]

Brian Kraznich?

Seriously, since the question was "Who", and I presume this a singular, that title would undoubtedly be BK.

Look at IcaLake now,

And imagine if all of your computer already had Icelake since 2017, and Intel already shipped their first 7nm EUV Mobile Chip with next gen uArch and GFx in 2019, and is shipping 7mn HP Desktop and Server next gen uArch as we speak. Also remember the original Intel 7nm is actually closer to TSMC 3nm then 5nm in density.

Intel would still have IPC, Node, along with Software Advantage if they didn't mess up their AVX segmentation. Imagine you had all that, and you saw Zen 2, and Ryzen 4000, would you still be hyped?

 

[Kenmitch]

I would say Intel and 10nm node.
Because if all was well with Intel, AMD would still not be competitive.
Just look at Nvidia and AMD.

At least your consistent, but in a negative way.

I'm not really sure why you even participate in any AMD related topics in this forum. It seems like you go out of your way to stir up the pot.

As far as the topic goes I'm guessing no one man/woman is responsible for AMD's recent success. In the end it's more or less a combination of events, decisions, efforts, goals, execution, focus, etc of the whole AMD team. I guess there's also a honorable mention to those who've purchased their products.

 

[Jaskalas]

"Saving AMD" is a result of success. Success is a result of competitive performance, and it takes two to dance that tango.

Intel had clearly missed a step, and AMD was in a healthy enough position to pounce. To seize the opportunity. As a consumer I am glad to see the competition. As others have said, "who" is a multiple answer question.

 

[AnandThenMan]

And, no, AMD is not a threat to NVIDIA. It may look so for some people who love AMD to no end, but it's just an illusion.

Never say never, remember when it was impossible AMD would ever, ever overtake Intel? Go back 4 years and claim AMD will have a 64 core monster that decimates Intel you would have been ridiculed off the internets.

Meanwhile people choose not to talk about their worst decision ever - which was to buy ATI. When did it happen? 15 years ago? They still don't have a single product which justified that purchase. The APU (the way it was initially envisaged) has never happened. This purchase almost tanked AMD.

AMD is going to be rolling out some very competitive (to put it mildly) mobile parts.

 

[VirtualLarry]

Yes, whipping up multi-die CPU packages to counter AMD is one of the oldest things in Intel's playbook... and the fact that they didn't immediately counter Zen by just slapping two Kaby Lake dies onto a package to instantly make an 8C CPU indicates that Skylake and its offshoots don't have the necessary on-die logic to do so, and/or there are other issues that make it infeasible.

Yes, Intel finally got rid of "FSB" (GTL+ QDR FSB).

I thought that was considered "progress", no? Could it have been a mistake? Could we possibly have had an alternate universe, in which we could drop a Skylake CPU into an i865G mobo, with DDR (1) RAM? LOL. Oh, but we could have gone with double-dies and double the cores, on that i865G mobo...

 

[amd6502]

Great management (Su, Papermaster) and great engineers, Raja, Moore, Keller, Plummer, Clark, and all the ones I don't know about.

Rory did good but Ruiz set up AMD for some financial hard times.

Boy I can't wait to get my hands on a Zen3 mobile laptop next year or so.

 

[DigDog]

AMD CEO Lisa Su is an actual engineer. He doctorate was in silicon-on-insulate mosfet design. She studied at MIT.

Intel CEO Bob Swan has a bachelor's in business management from U Buffalo.

 

[CHADBOGA]

AMD CEO Lisa Su is an actual engineer. He doctorate was in silicon-on-insulate mosfet design. She studied at MIT.

Intel CEO Bob Swan has a bachelor's in business management from U Buffalo.

Yeah but it was Krzanich who went to sleep whilst Intel's 10 process went down the gurgler, and Krzanich was a process guy. Go figure.

 

[AnandThenMan]

I know it sounds comical but Ruiz did save AMD, intentionally or accidentally. If AMD had kept fab ownership they would have sank into oblivion been bought out or closed shop.

But yes it is Dr. Su that has made the company legit she is easily one of the best CEOs out there.

 

[JasonLD]

AMD CEO Lisa Su is an actual engineer. He doctorate was in silicon-on-insulate mosfet design. She studied at MIT.

Intel CEO Bob Swan has a bachelor's in business management from U Buffalo.

Hector Ruiz and Brian Krzanich were also engineers. Having a engineering background certainly helps, but doesn't mean he/she will be a great CEO at a tech company by default. It also doesn't mean having a non-tech background would make a bad CEO at a tech company either.

 

[DigDog]

eh .. im gonna go with the gut feeling that having someone at the helm who has a history of processor design is far, far better than someone who is a businessman.

because i've met businessmen.

i'm not saying that having a degree guarantees you're not a moron. A lot of people who have a degree prove that they are resilient to boredom, not that they are smart.

But, to simplify, if a company that designs cars puts at the head of design someone who really, really love cars. Do you think you'd wind up with FIAT, or with Ferrari ?

 

[Ichinisan]

I think it was actually Intel that saved saved AMD -- by becoming complacent and failing to innovate (and having some nasty vulnerabilities with performance-crushing mitigation patches in OS).

 

[JoeRambo]

WIthout doubt Mr. Brian Krzanich did the most to ensure current AMDs recovery and success. Imagine alternate reality, where in 2015-16 Intel release 10nm CPUs on process with density and performance comparable or better to current TSMC 7nm. It is well possible alternate reality as well, cause Intel had substantial process lead for like 10 years at that point and TSMC 7nm does not require EUV or magic and could have been done in 2015-16.

It takes epic management blunders to squander lead and CEO always takes responsibility for high level failures.

 

[DrMrLordX]

Yeah but it was Krzanich who went to sleep

With a subordinate, no less!

 

[uzzi38]

And, no, AMD is not a threat to NVIDIA. It may look so for some people who love AMD to no end, but it's just an illusion.

AMD are most definitely a threat, but certainly not the biggest threat to Nvidia. Jensen's ego wins that crown.
Well, if you don't want to consider AMD a threat to Nvidia just now, that's fine. It's misguided, but overall fine. I look forwards to when RDNA2 launches and both Reddit and here implode at the notion that AMD can catch up significantly at all, then rely on yearly cadence to aim for the crown.

As an aside - you can make commments about process node and power efficiency, but why does it matter if AMD only match in power efficiency with a node jump? You're treating it like node jumps are easy to do, that they're just free performance or something. Staying a node ahead is an entirely valid way of maintaining a performance lead - and requires a significant amount of engineering and R&D to accomplish.

 

[caswow]

AMD are most definitely a threat, but certainly not the biggest threat to Nvidia. Jensen's ego wins that crown.
Well, if you don't want to consider AMD a threat to Nvidia just now, that's fine. It's misguided, but overall fine. I look forwards to when RDNA2 launches and both Reddit and here implode at the notion that AMD can catch up significantly at all, then rely on yearly cadence to aim for the crown.

As an aside - you can make commments about process node and power efficiency, but why does it matter if AMD only match in power efficiency with a node jump? You're treating it like node jumps are easy to do, that they're just free performance or something. Staying a node ahead is an entirely valid way of maintaining a performance lead - and requires a significant amount of engineering and R&D to accomplish.

People didnt give a s--- when intel was far ahead in process tech. now its "hurr durr amd needs 7nm to compete"

 

[Kenmitch]

Dang....Ran out of cyber popcorn!